Thanks for sending an issue! Please fill in the following template to help quickly solve your problem.

Document Link（文档链接）

https://gitcode.com/cann/ops-ras/blob/master/common/stub/inc/framework/stub_ops.h#L693-L699

Issues Section（问题文档片段）

* @li Atlas Inference Series Product in the scenarios where the data type is converted from FLOAT32 to INT8: \n
*     It can only guarantee that the input data has no precision errors within the range of (-2048, 1920).
* @li Atlas Inference Series Product in the scenarios where the data type is converted from FLOAT32 to INT64 and from FLOAT32 to UINT8: \n
*     It can only guarantee that the input data has no precision errors within the range of (-2147483648, 2147483583).
* @li Atlas Inference Series Product in the scenarios where the data type is converted from INT64 to FLOAT32: \n
*     It can only guarantee that the input data has no precision errors within the range of (-2147483648, 2147483647).
* @li Ascend 910_95 AI Processor in the scenario where the data type is converted from INT32 to INT4: \n
*     The last dim of x must be an even number.

Existing Issues（存在的问题）

The product names used in the Cast operator constraints comments of common/stub/inc/framework/stub_ops.h are outdated and inconsistent with the current official product naming conventions:

• "Atlas Inference Series Product" should be "Atlas inference products" (3 occurrences, lines 693/695/697).
• "Ascend 910_95 AI Processor" should be "Ascend 950PR/Ascend 950DT" (1 occurrence, line 699).

The outdated names may confuse users about which products the precision constraints apply to, and could propagate further if the header is used as a reference for other documentation or code generation.